Thach

Diamond and Related Materials 12 (2003) 799–805

Feasibility studies of magnetic particle-embedded carbon nanotubes for perpendicular recording media Cheng Tzu Kuo*, Chao Hsun Lin, An Ya Lo Department of Materials Science and Engineering, National Chiao Tung University, 1001 Ta-Hsueh Road, Hsinchu 300, Taiwan, ROC

Abstract Nano-sized magnetic particles were successfully used as the catalysts to synthesize magnetic metal-encapsulated carbon nanotubes (CNTs) or nanoparticles on Si wafers in a microwave plasma electron cyclotron resonance chemical vapor deposition (ECR-CVD) system with CH4 andyor H2 as source gases. The magnetic catalyst materials, including Fe–Pt, Co–Pt, Nd2Fe14B, Fe and Fe–Ni, were first deposited on Si wafers by a physical vapor deposition (PVD) method, with subsequent plasma treatment for nanoparticle transformation. The main process parameters include catalyst materials, hydrogen plasma catalyst pretreatment and deposition temperature. For applications in magnetic media, the process has the following advantages: perpendicularly aligned CNTs or nanoparticles; tip-growth CNTs; well-distributed magnetic particles; detectable magnetic field in each particle; high tube number density (up to 134 Gtubesyinch2 for Fe-assisted CNTs); favorable catalyst size; higher shape and induced anisotropy; and nanostructures that can be manipulated. The catalyst particle sizes of Fe, Nb2 Fe14 B and Fe–Pt (35–40 nm in diameter) are uniform and greater than but close to the critical optimum size or single domain size, which favor a higher coercive force. The greatest coercive force can reach 750 Oe for Fe-assisted CNTs at a deposition temperature of 715 8C, which is comparable with values reported in the literature. The coercive force difference between the vertical and horizontal directions can reach 300 Oe for Fe-assisted CNTs, and 355 Oe for Nb2Fe14B-assisted CNTs. 䊚 2002 Elsevier Science B.V. All rights reserved. Keywords: Carbon nanotubes; Catalyst; Cyclotron resonance CVD; Magnetic recording; Perpendicular recording media

1. Introduction Higher-areal-density magnetic recording media are being demanded in industries due to their great market volume. The way to increase the density is to minimize the signal spot size. Storage media can be broadly divided into three series, i.e. magnetic, optical and semiconductor storage media. Magnetic storage media include the traditional tapes and videocassettes, hard disks for mainframe computers, floppy disks for personal computers (PC), portable ZIP and MO disks (magnetooptic) with high storage capacity. Their development has covered a wide historical period, ranging from the industrial era to the mainframe and PC eras and up to the post-PC era, or in terms of years from 1970 to 2001. Because of the great market volume and the long development period, magnetic storage media have *Corresponding author. Tel.: q886-3-5731949; fax: q886-35721065. E-mail address: [email protected] (C.T. Kuo). 0925-9635/03/$ 䊚 2002 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 5 - 9 6 3 5 Ž 0 2 . 0 0 2 3 1 - 5

become an everyday commodity. For storage capacity, a minimum signal spot of -100 nm is required for a terabyte capacity, and of -20 nm for greater than terabyte capacity. Based on the orientation of the magnetic domains in magnetic storage media, they can be broadly divided into two categories, i.e. the horizontally oriented and vertically oriented types. The results reported indicate that horizontally-oriented magnetic storage media may reach the limit in physics for storage capacity of )40 Gbityinch2 w1,2x. In the present commercial hard-disk market, the horizontally oriented type is generally adopted for all magnetic storage media, except MO media, for which the vertically oriented type is used. In order to overcome the bottleneck of the horizontally oriented storage media, changing the recording type, such as vertically oriented or patterned media, is theoretically possible. A few hard disk manufacturers and research institutes have successfully designed such prototype media to demonstrate their feasibility w1,3–9x. However,

C.T. Kuo et al. / Diamond and Related Materials 12 (2003) 799–805

800

Table 1 Specimen designation, pretreatment and deposition conditions Specimen designation

Catalyst

Substrate temperature (8C)

Working pressure (10y3 torr)

H plasma etching

A1 A2 B1 B2 C1 C2 C3 C4 C5 C6 D1 D2 E1

Fe–Pt Fe–Pt Co–Pt Co–Pt Nd2Fe14B Nd2Fe14B Nd2Fe14B Nd2Fe14B Nd2Fe14B Nd2Fe14B Fe Fe Fe–Ni

705 673 715 673 717 590 620 660 685 700 715 673 705

6.2 5.5 6.3 5.5 6.4 6.2 5.8 6.6 5.9 6.5 6.3 5.5 6.2

Yes No Yes No Yes Yes Yes Yes Yes Yes Yes No Yes

wafers by electron cyclotron resonance chemical vapor deposition (ECR-CVD) methods. The catalyst magnetic materials, including Fe, Fe–Pt, Fe–Ni, Co–Pt and Nb2Fe14B, were coated on Si wafers with pure Fe or alloys as targets by PVD. Most of the magnetic metalcoated substrates were pretreated with hydrogen plasma under additional electrical heating to the desired temperature. After hydrogen plasma pretreatment, the specimens immediately underwent CNT deposition. The effects of magnetic catalyst materials, pretreatments and deposition temperature were examined. The specimen designation, catalyst pretreatment and deposition conditions are shown in Table 1. Magnetic field strength of 875 G was applied to maintain the ECR condition. The magnetic properties, morphology, microstructure and bonding structures of the magnetic catalyst-assisted carbon nanostructures were characterized by vibration sample magnetometry (VSM), magnetic force microscopy (MFM), atomic force microscopy (AFM), scanning electron microscopy (SEM), transmission electron microscopy (TEM) and high-resolution transmission electron microscopy (HRTEM).

Other deposition conditions: microwave power, 800 W; magnetic field, 875 G; flow ratio, CH4 yH2 s15 sccm:15 sccm; and bias, y200 V. H plasma etching conditions: H2, 15 sccm; bias, y200 V.

many problems must be solved before commercialization, including: (1) design and fabrication techniques for a high-precision recording head and servo driving system; (2) fabrication parameters for storage media, such as size uniformity (dispersion -10%), morphology control, higher magnetic properties (higher coercive field strength Hc, higher squareness ratio S, higher anisotropic magnetic crystal, lower noise, etc.); lower cost, etc. This work was planned to study the feasibility of nano-resolution magnetic storage media using state-ofthe-art nanostructure fabricating techniques. Magnetic nanoparticles on substrates act as the catalysts to synthesize protective layers of carbon nanotubes or nanoparticles on themselves.

3. Results and discussion 3.1. Effects of magnetic catalyst material The morphology features, adhesion and magnetic properties of the magnetic metal-assisted CNTs are shown in Table 2. The corresponding side-view and topview SEM micrographs for different magnetic catalysts are depicted in Fig. 1 for Fe, Fe–Pt, Co–Pt, Nb2Fe14B and Fe–Ni catalysts. From the point of view of applications in recording media, all CNTs here take advantage of the vertically aligned and tip-growth CNTs instead of base-growth CNTs. The results show that under the same deposition conditions, different catalysts can produce CNTs with different morphology features and properties, such as differences in tube number density, tube length, carbon film formation, bonding between catalyst and CNTs, growth mechanism and type of

2. Experimental Carbon nanotubesynanoparticles were synthesized on magnetic metal-coated substrates of p-type silicon Table 2 Morphology features and properties of the catalyst-assisted CNTs Specimen designation

Catalyst material Tube length (nm) Tube diameter (nm) Tube number density (Gtubesyinch2) Diameter of catalyst particle (nm) Graphene layer alignment Adhesion Tube alignment Tube type

A1

B1

C1

D1

E1

Fe–Pt 170 50 77 40 Poor Fair Very high Hollow

Co–Pt 170 60–180 16 20–120 – Fair Very high Hollow

Nd2Fe14B 2100 60 71 35 Good Fair High Bamboo

Fe 1500 60 134 35 – Poor High Bamboo

Fe–Ni 600 50–120 39 10–100 Good Poor High Bamboo

Deposition conditions: CH4 yH2s15 sccm:15 sccm; microwave power, 800 W; bias, y200 V; and magnetic field, 875 G.

C.T. Kuo et al. / Diamond and Related Materials 12 (2003) 799–805

801

Fig. 1. SEM micrographs of the different catalyst-assisted CNTs: (a) Fe, side view; (b) Fe, top view (specimen D1); (c) Fe–Pt, side view; (d) Fe–Pt, top view (specimen A1); (e) Co–Pt, side view; (f) Co–Pt, top view (specimen B1); (g) Nd2Fe14B, side view; (h) Nd2Fe14B, top view (specimen C1); (i) Fe–Ni, side view; and (j) Fe–Ni, top view (specimen E1).

802

C.T. Kuo et al. / Diamond and Related Materials 12 (2003) 799–805

Fig. 2. SEM micrographs of the different catalyst-encapsulated carbon nano-particles without the catalyst H plasma pretreatment: (a) Fe (specimen D2); (b) Fe–Pt (specimen A2); and (c) Co–Pt (specimen B2).

CNTs. These differences in structure or properties may relate to the solubility difference of carbon in catalysts, interaction of the catalyst with the substrate, and the etching rate difference between CNTs and carbon films in hydrogen plasma. For example, the carbon solubility of Fe is much better than Fe–Pt, and Co is much better than Co–Pt, due to the very limited carbon solubility of non-ferrous metals, such as Pt. Under the present conditions, the maximum tube number density reached 134 Gtubesyinch2 for Fe-assisted CNTs. The tube number density is one of the main factors determining the media recording density. The longest tube length can reach 2100 nm for 15 min of deposition time for Nd2Fe14Bassisted CNTs, which approximately corresponds to the highest growth rate. The highest coercive force is ;750 Oe for the Fe-assisted CNTs. It is interesting to note that bamboo-like CNTs can be formed without using nitrogen gas, although nitrogen was proposed as one of the main parameters for the formation of bamboo-like CNTs w10,11x. For certain applications, if the removal

of catalysts from the tips of CNTs is required, this can easily be achieved by selecting a suitable catalyst combined with ultrasonic agitation in an acetone bath, e.g. Fe–Pt and Co–Pt catalysts.

Fig. 3. Lattice image of the Nd2Fe14B-assisted CNTs close to the catalyst particle area (specimen C1).

Fig. 4. HRTEM image of the Fe-assisted CNTs close to the catalyst particle area (specimen D1).

3.2. Effects of hydrogen plasma pretreatment SEM micrographs for the magnetic metal-assisted CNTs or nanoparticles show differences in the catalysts with and without hydrogen plasma pretreatments by comparing Fig. 1a with Fig. 2a, Fig. 1c with Fig. 2b, and Fig. 1e with Fig. 2c for Fe, Fe–Pt and Co–Pt catalysts, respectively. It is obvious that the morphology of the carbon nanostructures changes from tube-like to particle-like shapes on omitting the hydrogen plasma pretreatment step. From the point of view of applications in recording media, the particle-like shapes have the advantage of being short and rigid.

C.T. Kuo et al. / Diamond and Related Materials 12 (2003) 799–805

803

Table 3 The magnetic properties and structure features of the magnetic catalyst-assisted CNTs Specimen Catalyst designation material

Diameter Tube number Hc,V Hc,H of catalyst density (Oe) (Oe) particle (nm) (Gtubesyinch2)

A1 B1 C1 D1 E1

40 20–120 35 35 10–100

Fe–Pt Co–Pt Nd2Fe14B Fe Fe–Ni

77 16 71 134 39

– – 705 750 300

– – 350 450 300

Hc,V, coercive force in the vertical direction of the substrate surface; Hc,H, coercive force in the horizontal direction of the substrate surface.

3.3. Catalyst shape Typical TEM and HRTEM lattice images of the Nd2Fe14B- and Fe-assisted CNTs close to the catalyst particles are shown in Figs. 3 and 4, respectively. In addition to the advantage of the vertically aligned CNTs, the catalysts at the tips of the CNTs are pear-like in shape, which favor a greater contribution from magnetic shape anisotropy. The present process under ECR conditions with magnetic field strength of 875 G may also take the advantage of a magnetic annealing effect to enhance the induced anisotropy. The magnetic properties and structure features of the magnetic catalyst-assisted CNTs are shown in Table 3. The catalyst particles show a wider size distribution, from 10 to 120 nm in diameter, for the Co–Pt and Fe–Ni catalysts. In contrast, the catalyst particles are more uniform in size, approximately 35–40 nm in diameter, for the Fe, Nb2Fe14B and Fe– Pt catalysts. It seems that the uniform particle size results in greater coercive force and magnetic anisotropy. This may relate to a size greater than but close to the critical optimum size or single domain size, which favors higher coercive force w11,12x. The greatest coercive force can reach 750 Oe for Fe-assisted CNTs at a deposition temperature of 715 8C, which is comparable with values reported in the literature. The difference in coercive force between the vertical and horizontal directions can reach 300 Oe for Fe-assisted CNTs, and 355 Oe for Nb2Fe14B-assisted CNTs.

Fig. 5. Two typical hysteresis loops of the Nd2Fe14B-assisted CNTs in both vertical and horizontal directions to the substrate surface (specimen C6).

high enough temperatures, the aspect ratio of the catalyst particle may also be increased during CNT growth by reducing the viscosity of the catalyst. 3.5. AFM and MFM images The AFM and the corresponding MFM images and MFM line-scan profiles for the Fe-, Fe–Pt- and Co–Ptassisted nanoparticles or CNTs are shown in Fig. 7a–c, d–f and g–i, respectively. The brighter and darker areas in the MFM images represent areas with repulsive and attractive magnetic forces to the MFM probe, respectively. This implies that the magnetic particles are well distributed, and the magnetic field of each particle can be detected or read. 4. Conclusions Five different magnetic materials were successfully used to act as catalysts for the synthesis of magnetic

3.4. Effect of deposition temperature The two typical hysteresis loops of the Nd2Fe14Bassisted CNTs in the vertical and horizontal directions of the substrate surface are shown in Fig. 5. The coercive force of the Nd2Fe14B-assisted CNTs as a function of deposition temperature is shown in Fig. 6. It is obvious that the higher deposition temperature favors greater coercive force and greater anisotropy. The abnormally high anisotropy at temperatures greater than 700 8C is not understood. It may relate to temperatures high enough to enhance the magnetic annealing effect. At

Fig. 6. Coercive forces in the vertical and horizontal directions vs. deposition temperature for the Nd2Fe14B-assisted CNTs (specimens C1–C6).

804

C.T. Kuo et al. / Diamond and Related Materials 12 (2003) 799–805

Fig. 7. AFM image, MFM image and the corresponding MFM line scan profile for different catalyst-assisted CNTs or nano-particles, respectively: (a,b,c) Fe (specimen D2); (d,e,f) Fe–Pt (specimen A1); (g,h,i) Co–Pt (specimen B2).

metal-encapsulated CNTs or nanoparticles on Si wafer to study their feasibility for perpendicular recordingmedia applications. The results show that we can successfully deposit perpendicularly-aligned magnetic particle-embedded nanotubes and nanoparticles. For recording media applications, most of the deposited CNTs take advantage of tip-growth and perpendicularly aligned CNTs. CNTs with different tube number density, tube length, carbon film formation, bonding between catalyst and CNTs, growth mechanism and type of CNT

can be manipulated by selecting different catalysts and pretreatments. These differences in structure or properties may relate to the different solubility of carbon in the catalysts, interaction of the catalyst with the substrate and the etching rate difference between CNTs and carbon films in hydrogen plasma. The tube number density is a key parameter determining the density of the recording media. Under the present conditions, the maximum tube number density can reach 134 Gtubesy inch2 for Fe-assisted CNTs.

C.T. Kuo et al. / Diamond and Related Materials 12 (2003) 799–805

With regard to the magnetic properties of the magnetic metal-encapsulated carbon nanostructures, the grain size of the magnetic particles (35–40 nm in diameter) is greater than but close to the critical optimum size or single domain size, which favors a higher coercive force. A higher CNT deposition temperature results in a greater coercive force due to smaller catalyst size; the greatest coercive force can reach 750 Oe for Fe-assisted CNTs at a deposition temperature of 715 8C, which is comparable to values reported in the literature w12–15x. The process also takes advantages of higher shape and induced anisotropy due to its higher aspect ratio and magnetic annealing effect. The coercive force difference between the vertical and horizontal directions can reach 355 Oe under the present conditions. The results also demonstrate the potential applications in magnetic recording media of the isolated and well-distributed magnetic particles in the magnetic metal-encapsulated carbon nanostructures as imaged by MFM micrographs. Acknowledgments This work was supported by the National Science Council (Contract No NSC90-2216-E-009-040, -034 and -035) of Taiwan.

805

References w1x M. Todorovic, S. Schultz, J. Wong, A. Scherer, Appl. Phys. Lett. 74 (1999) 2516. w2x D.E. Speliotis, J. Magn. Magn. Mater. 193 (1999) 29. w3x S.Y. Chou, Proc. IEEE 85 (1997) 652. w4x S.Y. Chou, P.R. Krauss, J. Magn. Magn. Mater. 155 (1996) 151. w5x S.Y. Chou, P.R. Krauss, L. Kong, J. Appl. Phys. 79 (8) (1996) 6101. w6x S. Iwasaki, K. Takemura, IEEE Trans. Magn. 11 (5) (1975) 1173. w7x S. Iwasaki, Y. Nakamura, IEEE Trans. Magn. 13 (5) (1977) 1272. w8x R.L. White, J. Magn. Magn. Mater. 209 (2000) 1. w9x D.L. Leslie-Pelecky, R.D. Rieke, Chem. Mater. 8 (1996) 1770. w10x X. Ma, E. Wang, D.A. Jefferson, et al., Appl. Phys. Lett. 75 (1999) 3105. w11x H.L. Chang, C.H. Lin, C.T. Kuo, Diamond Relat. Mater. 11 (3–6) (2002) 793. w12x X.X. Zhang, G.H. Wen, S. Huang, L. Dai, R. Gao, Z.L. Wang, J. Magn. Magn. Mater. 231 (2001) L9. w13x N. Grobert, W.K. Hsu, Y.Q. Zhu, et al., Appl. Phys. Lett. 75 (1999) 3363. w14x X. Sun, G.A. Yacaman, M. Jose, D. Xinglong, J. Shouri, Mater. Sci. Eng. A 286 (2000) 157. w15x T. Hayashi, S. Hirono, M. Tomita, S. Umemura, Nature 381 (1996) 772.